Setup & Installation
Or with the ClawHub CLI, for registry-managed skill folders outside a full OpenClaw workspace:
Version History
fix warnings
What This Skill Does
Routes coding tasks to opencode, a CLI-based AI coding agent. The skill manages a two-phase loop: planning first, then building, never skipping either. It passes user instructions to opencode and interprets results to decide what comes next.
Enforces a plan-before-build discipline that prevents unreviewed code generation, reducing implementation errors from misunderstood requirements.
When to use it
- Planning a new feature before writing any code
- Implementing an approved plan in a repository
- Debugging a codebase by delegating analysis to opencode
- Iterating through plan-build cycles until requirements are met
- Answering technical questions about a repo without manual investigation
Example Workflow
Here's how your AI assistant might use this skill in practice.
User asks: Add a user authentication module to this repo
- 1Run opencode with the plan agent: opencode run "Add a user authentication module" --continue --agent plan
- 2Review the step-by-step plan returned by opencode
- 3If the plan is incomplete, ask opencode to revise it
- 4Once the plan is approved, run opencode with the build agent to implement it
- 5If opencode asks a question during build, switch back to plan agent to clarify
Authentication module implemented according to the reviewed and approved plan
Requirements
Accounts, API keys, or tools you or your AI assistant may need to set up while using this skill.